In a method for fabricating a self-aligned vertical comb drive structure, a multi-layer structure is first formed. The multi-layer structure includes inter-digitated first and second comb structures formed via etching using a first mask layer as a mask. The first comb structure includes a plurality of first comb fingers, each having a first finger portion formed in a first device layer and a second finger portion formed in a second device layer and separated from the first finger portion by a self-aligned pattern on a stop layer. The second comb structure includes a plurality of second comb fingers formed solely in the second device layer. The second finger portions of the first comb fingers are subsequently removed.